OpenAI – Show me the Money!

Atlas is an attempt to monetise the 700m.

• OpenAI has launched a browser, which is clearly a further pivot towards becoming a consumer ecosystem in its own right, and if successful, will solve the problem of the 700m+ users who don’t pay it a cent.
• OpenAI’s single biggest problem is not the lack of compute capacity but the fact that the vast majority of its user base does not pay it any money at all.
• This is exacerbated by the fact that a search or a query on ChatGPT is vastly more expensive than one run on Google, which I think is the main reason why OpenAI is burning billions of dollars every quarter.
• RFM Research concluded that a business like ChatGPT should become cash generative when it hits 10s of millions of users, even with a price of way less than $20 per month.
• However, because these users pay $0 and there are 700m or more of them, what should be a cash-generative business rapidly becomes a financial millstone.
• Consequently, one of the main issues that OpenAI faces in its quest to become a going concern is how to monetise the 700m users.
• I have long expected that this would be achieved through some form of advertising model, but with ChatGPT’s current format, this is problematic.
• If, however, OpenAI can entice its users to use the Atlas browser, it will remove them from Google’s orbit as the default search provider and monetisation machine into its own.
• At the same time, with users spending more time in an OpenAI asset, its opportunities for monetisation via advertising will become much greater.
• If it can generate $6 per user per month, this will generate $50.4bn in revenues, which will go a long way to offsetting the cost of delivering the service to all of its users and probably its training budget as well.
• OpenAI clearly intends to create a new digital ecosystem complete with APIs, developers and an app store (for AI services), which is how it could use the dislocation caused by AI to disrupt the consumer digital ecosystem dominated by Apple, Google and Tencent.
• Market dislocation is always an opportunity for market leadership to change, and the advent of generative AI is potentially highly disruptive to the existing digital ecosystem.
• If most of the digital life services that users enjoy become controlled by a digital agent, then the point of control in the digital ecosystem will have shifted once again, meaning that it will be vital to be the creator and owner of the agent that users interface with.
• Google is the company that looks the most susceptible to this disruption, but it already has an agent of its own and is the clear no. 2 in the market.
• Furthermore, with over 3bn users of at least one Google service, it has the opportunity to entice those users to use its agent rather than the one of OpenAI and thereby maintain control if this shift happens.
• Apple, on the other hand, looks much safer in the short term, but as it does not have an agent with anything like the capabilities of ChatGPT or Gemini, it could be in real danger of commoditisation if this trend materialises before Apple has something ready for its users.
• This is how OpenAI could start to justify its $500bn valuation, but it has a very long way to go before its ecosystem is successful and will face stiff resistance along the way.
• Furthermore, it has to accomplish this before there is a correction in the hype bubble that is currently providing it with almost free money and driving an unprecedented roll-out of compute capacity across the globe.
• This correction could happen at any time, and it is at that point that everyone will begin to question how you make a return on a data centre that costs $25bn to build.
• I still think that the correction will catch OpenAI before it becomes financially viable, and as such, I see real trouble ahead for those that are not already generating cash.
• Hence, I continue to see the possibility of OpenAI being acquired by one of the larger players (possibly Apple), Anthropic being acquired by Amazon (badly needed), as well as a large round of consolidation, bankruptcies and exits.
• It is the companies that are making money from the rollout that will fare the best as they have revenues and profits now and are far more fairly valued.
• That being said, I still prefer my adjacencies of inference at the edge and nuclear power, where I still have positions in my portfolio.